Back to Projects
AI/MLFeatured
AI-Enabled SQL Query Agent
Natural language to SQL for network incident management
AI/Backend Developer
Aug 2024 — Dec 2024
01. Overview
An intelligent agent that converts natural language questions into optimized SQL queries for network incident analysis and real-time device monitoring.
💡 Impact: Enabled non-technical users to query complex network data using plain English
02. The Problem
Network operations teams needed to query incident data without writing SQL, while maintaining security and query optimization for time-series data.
03. Approach
Built a multi-node StateGraph workflow using LangGraph with intent recognition, entity resolution, and domain-specific SQL generation with injection protection.
04. Tech Stack
PythonLangGraphLangChainOpenAI GPTPostgreSQLTimescaleDBDocker
05. Results & Impact
- ✓Achieved 95% accuracy on supported query patterns with intelligent fallback handling
- ✓Reduced average query time for ops teams from minutes to seconds
- ✓Deployed secure NL-to-SQL pipeline for a leading Network-as-a-Service provider